In Perl_ck_subr(), no need to create/set o3 in the !proto block.

A good optimising compiler can already spot this, but removing dead code makes
it easier for the humans.

1 files changed, 2 insertions, 5 deletions

diff --git a/op.c b/op.c
index b27c2ea013..8aa1cae581 100644
--- a/op.c
+++ b/op.c
@@ -8450,18 +8450,15 @@ Perl_ck_subr(pTHX_ OP *o)
 
     if (!proto) {
 	while (o2 != cvop) {
-	    OP* o3;
 	    if (PL_madskills && o2->op_type == OP_STUB) {
 		o2 = o2->op_sibling;
 		continue;
 	    }
-	    if (PL_madskills && o2->op_type == OP_NULL)
-		o3 = ((UNOP*)o2)->op_first;
-	    else
-		o3 = o2;
+
 	    /* Yes, this while loop is duplicated. But it's a lot clearer
 	       to see what is going on without that massive switch(*proto)
 	       block just here.  */
+
 	    list(o2); /* This is only called if !proto  */
 
 	    mod(o2, OP_ENTERSUB);